A media fill can look finished on paper and still tell you very little.
The real question is not whether the exercise was completed.
The real question is whether it behaved like the process you are asking it to represent.
That distinction matters because sterile manufacturing does not fail in the clean, convenient version of the process. It fails at the edges. During interventions. During awkward setups. During the steps people know are difficult and quietly learn to work around.
In a September 2026 warning letter, FDA described problems with an aseptic manufacturer's process simulations and investigations. Among the issues FDA identified: media fills that did not adequately represent maximum intervention counts, inherent interventions that were not documented in batch records, and media-fill batch sizes that were insufficient to assess the actual contamination risk.
FDA also challenged the investigation logic. The firm attributed several media-fill failures to deficient operator aseptic technique. But the investigation did not ask how those practices went undetected by the aseptic observer program during commercial operations.
That is the part worth sitting with.
The simulation has to carry the real risk
A process simulation is not a ceremonial rehearsal. It is a test of whether the process, the people, the equipment, the interventions, and the controls work together under representative conditions.
If the hardest interventions are left out, counted inaccurately, or documented loosely, the simulation is testing a cleaner version of reality.
That may be easier to pass.
It is also less useful.
The practical questions are straightforward:
- Are the maximum interventions represented?
- Are inherent interventions captured in the batch record, not just remembered afterward?
- Is the number of units large enough to support the conclusion?
- Does the simulation reflect the actual line layout and operator movement?
- Are the awkward, repetitive, and high-exposure steps included?
If the answer to those questions is unclear, the issue is not just a media-fill protocol. It is the quality of the assurance being created.
Operator error is not a root cause by itself
Operator technique can absolutely contribute to an aseptic failure.
But stopping there is often a way to end the investigation before the system has been examined.
When an operator makes the same kind of mistake more than once, the next questions should be uncomfortable and specific:
- Was the intervention physically difficult?
- Did the layout force awkward reach, contact, or repeated manipulation?
- Was the procedure usable in the moment, or only readable in a conference room?
- Did training match the actual task?
- Could an observer reliably see the behavior that mattered?
- Did the observer program have enough authority and consistency to detect drift?
- Were similar operators, lines, and interventions reviewed?
People work inside systems. If the system makes the right action difficult and the wrong action invisible, retraining alone is not a durable correction.
The observer program is part of the control strategy
An observer program is not background theater. It is one of the ways a site detects loss of aseptic discipline before the problem becomes a batch problem.
That means it needs more than a schedule and a completed form.
The site should know what observers are looking for, how observations are calibrated, how trends are escalated, and what happens when commercial behavior does not match training behavior.
The question is simple:
Would the observer program reliably catch the behavior that caused the media-fill failure if it happened on a normal production day?
If the answer is no, the media fill did its job. It exposed a weakness. The investigation has to do the rest.
The useful close
The point is not to make media fills harder for the sake of making them harder.
The point is to make them honest enough to help.
That means simulating the process you actually run, documenting the interventions you actually perform, and investigating the detection system when poor practices escape notice.
A passing simulation is reassuring only when the simulation earned that reassurance.
That is the work: fewer convenient assumptions, better evidence, and controls that still function when the process gets busy.